Nobiletin Inhibits Human Osteosarcoma Cells Metastasis by Blocking ERK and JNK-Mediated Mmps Expression

www.impactjournals.com/oncotarget/ Oncotarget, Vol. 7, No. 23

Nobiletin inhibits human osteosarcoma cells metastasis by blocking ERK and JNK-mediated MMPs expression

Hsin-Lin Cheng1, Ming-Ju Hsieh1,2, Jia-Sin Yang1,3, Chiao-Wen Lin4,5, Ko-Haung Lue6,7, Ko-Hsiu Lu6,8, Shun-Fa Yang1,3 1Institute of Medicine, Chung Shan Medical University, Taichung 40201, Taiwan 2Cancer Research Center, Changhua Christian Hospital, Changhua 500, Taiwan 3Department of Medical Research, Chung Shan Medical University Hospital, Taichung 40201, Taiwan 4Institute of Oral Sciences, Chung Shan Medical University, Taichung 40201, Taiwan 5Department of Dentistry, Chung Shan Medical University Hospital, Taichung 40201, Taiwan 6School of Medicine, Chung Shan Medical University, Taichung 40201, Taiwan 7Department of Pediatrics, Chung Shan Medical University Hospital, Taichung, Taiwan 8Department of Orthopedics, Chung Shan Medical University Hospital, Taichung 40201, Taiwan Correspondence to: Ko-Hsiu Lu, e-mail: [email protected] Shun-Fa Yang, e-mail: [email protected] Keywords: nobiletin, metastasis, MMP, CREB, SP-1 Received: December 01, 2015 Accepted: April 11, 2016 Published: April 29, 2016

ABSTRACT

Nobiletin, a polymethoxyflavone, has a few pharmacological activities, including anti-inflammation and anti-cancer effects. However, its effect on human osteosarcoma progression remains uninvestigated. Therefore, we examined the effectiveness of nobiletin against cellular metastasis of human osteosarcoma and the underlying mechanisms. Nobiletin, up to 100 μM without cytotoxicity, significantly decreased motility, migration and invasion as well as enzymatic activities, protein levels and mRNA expressions of matrix metalloproteinase (MMP)-2 and MMP-9 in U2OS and HOS cells. In addition to inhibition of extracellular signal-regulated kinase (ERK) and c-Jun N-terminal kinase (JNK), the inhibitory effect of nobiletin on the DNA- binding activity of the transcription factor nuclear factor-kappa B (NF-κB), cAMP response element-binding protein (CREB), and specificity protein 1 (SP-1) in U2OS and HOS cells. Co-treatment with ERK and JNK inhibitors and nobiletin further reduced U2OS cells migration and invasion. These results indicated that nobiletin inhibits human osteosarcoma U2OS and HOS cells motility, migration and invasion by down- regulating MMP-2 and MMP-9 expressions via ERK and JNK pathways and through the inactivation of downstream NF-κB, CREB, and SP-1. Nobiletin has the potential to serve as an anti-metastatic agent for treating osteosarcoma.

INTRODUCTION disease. Recently, the radical surgery and chemotherapy for osteosarcoma have advanced and improved survival Osteosarcoma is the most prevalent and high-graded rates to approximately 70% at 5 years, but the results form of primary bone cancer, comprising approximately remain unsatisfactory, particularly of those pertaining to 60% of all bone sarcomas [1, 2]. In 1970s, surgery alone cancer metastasis. cured less than 20% of patients with osteosarcoma because Metastasis is a complex process that includes the majority of deaths resulted from distant metastasis. several events collectively termed the invasion- Approximately 20% of patients with osteosarcoma will metastasis cascade [3]. The coordination of several relapse within 5 years and have detectable metastasis at signaling pathways results in the detachment of tumor presentation and the lung is the predominant site of distant cells, motility, degradation of the extracellular matrix

www.impactjournals.com/oncotarget 35208 Oncotarget (ECM), invasion, migration, adhesion to endothelial RESULTS cells, and the reestablishment of growth at a distant site [4]. Initially, the epithelial-mesenchymal transition Cytotoxicity of nobiletin on U2OS, HOS and (EMT) combines loss of epithelial cell junction proteins MC3T3-E1 cells (e.g. E-cadherin) and the gain of mesenchymal markers, such as N-cadherin and Vimentin [5]. Degradation To determine the cytotoxic effects of nobiletin on of ECM by proteolytic enzymes, such as matrix human osteosarcoma cells and mouse osteoblast cells, metalloproteinases (MMPs), plays a critical role the effects of serially diluted nobiletin on U2OS, HOS in angiogenesis, tumor formation and progression, and MC3T3-E1 cells were examined using the MTT and microenvironment formation. ECM degradation assay. Nobiletin (0-100 μM) did not effectively inhibit by cancer cells through MMPs may disrupt the the cell viability of U2OS (p = 0.083), HOS (p = 0.888) intercellular matrix to promote cancer cell mobility, and MC3T3-E1 (p = 0.972) cells after treatment for 24 h eventually causing metastasis [6–8]. Among MMPs, (Figure 1B). Therefore, we used this concentration range MMP-2 (gelatinase A) and MMP-9 (gelatinase B) are for nobiletin in the subsequent experiments involving vital enzymes involved in the degradation of gelatin, osteosarcoma cells. collagen and laminin, main components of ECM [9] and high expressions of them are indicators of high Nobiletin inhibits cell invasion and migration of invasive potential of cancer cells and are related to poor U2OS and HOS cells prognosis of patients [3, 10, 11]. In contrast, down- regulation of MMP-2 and MMP-9 provides a preventive As shown in Figure 1C and 1D, nobiletin dose- and measure against tumor metastasis [12, 13]. Additionally, time-dependently reduced U2OS and HOS cells moving the 5' end of promoter region of the MMP-2 and MMP- into the wound (U2OS: p < 0.001; HOS: p < 0.001). 9 gene contains a variety of regulatory sequences that Similarly, using a modified chamber with or without binds well-characterized transcriptional units of MMP-2 Matrigel, nobiletin markedly reduced migration (U2OS: and MMP-9, including nuclear factor-kappa B (NF-κB), p < 0.001; HOS: p < 0.001) and invasion (U2OS: p < SP-1, cAMP response element-binding protein (CREB), 0.001; HOS: p < 0.001) activities in U2OS and HOS cells and activator protein (AP)-1, possibly contributing to (Figure 1E and 1F). The results suggested that nobiletin tumor metastasis [14, 15]. significantly inhibits the motility, migration potential and Recently, there is an increasing focus on naturally invaseness of U2OS and HOS cells. occurring plant products, that provides a preventive strategy for people with a high risk of cancers [16]. Nobiletin reduces expressions and proteolytic Flavonoids, a vast group of polyphenolic compounds activities of MMP-2 and MMP-9 of U2OS and ubiquitously present in plants [17], exhibit a wild HOS cells spectrum of pharmacological properties and offer significantly protection against various cancers [18, For clarifying whether MMP-2 and MMP-9 involves 19]. As a member of flavonoids, nobiletin (5, 6, 7, 8, the nobiletin-induced suppression of cell motility and 3’ 4’-hexamethoxyflavone) has been proposed to have invasion, gelatin zymography was used. It showed that anti-cancer properties, including anti-angiogenesis, nobiletin dose-dependently reduced enzyme activities of anti-proliferation, anti-metastasis and induced apoptosis MMP-2 (U2OS: p < 0.001; HOS: p < 0.001) and MMP-9 in vitro and in vivo [20–23]. Accumulate studies have (U2OS: p < 0.001; HOS: p < 0.001) in U2OS and HOS showed that nobiletin also suppressed the expression cells (Figure 2A). Also, western blotting revealed that of MMPs, which may serve as a promising agent nobiletin progressively reduced protein expressions of for suppression of cancer invasion and metastasis MMP-2 (U2OS: p < 0.001; HOS: p < 0.001) and MMP- [24]. Moreover, nobiletin considerably attenuates 9 (U2OS: p < 0.001; HOS: p < 0.001) (Figure 2B). To metastasis via both ERK and PI3K/Akt pathway in further clarify the down-regulatory effects of nobiletin on HGF-treated liver cancer HepG2 cells, indicating protease, reverse transcription polymerase chain reaction the greater bioavailability of nobiletin than apigenin, (PCR) and quantitative PCR were conducted. Actually, tricetin and tangeretin. Therefore, nobiletin shows nobiletin significantly reduced mRNA expressions of the higher potential of being developed for clinical MMP-2 (RT-PCR—U2OS: p < 0.001; HOS: p < 0.001. applications [25]. However, the effects of nobiletin q-RT-PCR—U2OS: p = 0.019; HOS: p < 0.001) and on human osteosarcoma migration and invasion, and MMP-9 (RT-PCR—U2OS: p < 0.001; HOS: p < 0.001. the underlying mechanisms are not yet defined. In the q-RT-PCR—U2OS: p < 0.001; HOS: p < 0.001) dose- present study, we investigated the potential inhibitory dependently in U2OS and HOS cells (Figure 2C and effects of nobiletin on osteosarcoma metastasis in vitro 2D). The results indicated that the nobiletin-induced and also revealed the possible molecular mechanisms. antimetastatic effect is related to the inhibition of the

www.impactjournals.com/oncotarget 35209 Oncotarget Figure 1: Nobiletin inhibits in vitro wound closure, migration and invasion in the U2OS and HOS cells. A. The structure of nobiletin. B. U2OS and HOS human osteosarcoma cells and MC3T3-E1 mouse osteoblast cells were treated with nobiletin (0-100 μM) in serum free medium for 24 h by MTT assay are illustrated. C. and D. U2OS and HOS cells were wounded and then treated with nobiletin (0–100 μM) for 24 h in a serum-containing medium. At 0 h, 6 h, 12 h and 24 h, phase-contrast pictures of the wounds at four different locations were taken. E. Cell migration was measured using a Boyden chamber for 24 h with polycarbonate filters. F. Cell invasion was measured using a Matrigel-coated Boyden chamber for 48 h with polycarbonate filters. Cytotoxicity effects: U2OS: F = 2.827, p = 0.083; HOS: F = 0.274, p = 0.888; MC3T3-E1: F = 0.121, p = 0.972. Concentration effects: wounding healing (U2OS: F = 279.878, p < 0.001; HOS: F = 109.236, p < 0.001); cell migration (U2OS: F = 296.344, p < 0.001; HOS: F = 464.672, p < 0.001); invasion (U2OS: F = 327.312, p < 0.001; HOS: F = 196.642, p < 0.001). aSignificantly different, p < 0.05, when compared with 0 μM. bSignificantly different, p < 0.05, when compared with 25 μM. cSignificantly different, p < 0.05, when compared with 50 μM. dSignificantly different, p < 0.05, when compared with 75 μM. Time effects: wound healing (U2OS: F = 1555.239, p < 0.001; HOS: F = 928.975, p < 0.001). †Significantly different, p < 0.05, when compared with the vehicle group. ‡Significantly different, p < 0.05, when compared with 6 h. #Significantly different, p < 0.05, when compared with 12 h. www.impactjournals.com/oncotarget 35210 Oncotarget Figure 2: Nobiletin inhibits MMP-2 and MMP-9 proteolytic activity, protein and mRNA expression. A. U2OS and HOS cells were treated with nobiletin (0–100 μM) for 24 h in a serum free medium and then subjected to gelatin zymography to analyze the activity of MMP-2 and MMP-9. B. Western blotting to analyze the protein levels of MMP-2 and MMP-9. Quantitative results of MMP-2 and MMP-9 protein levels, which were adjusted with β-actin protein level. C. RT-PCR and D. q-RT-PCR to analyze the mRNA expression of MMP-2 and MMP-9. Quantitative MMP-2 and MMP-9 mRNA levels were adjusted to GAPDH level. Concentration effects: gelatin zymography (MMP-2 of U2OS: F = 60.254, p < 0.001 and HOS: F = 160.282, p < 0.001; MMP-9 of U2OS: F = 133.253, p < 0.001 and HOS: F = 67.263, p < 0.001); western blot (MMP-2 of U2OS: F = 92.251, p < 0.001 and HOS: F = 25.151, p < 0.001; MMP-9 of U2OS: F = 25.156, p < 0.001 and HOS: F = 172.876, p < 0.001); RT-PCR (MMP-2 of U2OS: F = 68.584, p < 0.001 and HOS: F = 138.258, p < 0.001; MMP-9 of U2OS: F = 60.675, p < 0.001 and HOS: F = 63.82, p < 0.001); q-RT-PCR (MMP-2 of U2OS: F = 8.363, p = 0.019 and HOS: F = 116.834, p < 0.001; MMP-9 of U2OS: F = 186.999, p < 0.001 and HOS: F = 52.371, p < 0.001). aSignificantly different,p < 0.05, when compared with 0 μM. bSignificantly different,p < 0.05, when compared with 25 μM. cSignificantly different,p < 0.05, when compared with 50 μM. dSignificantly different, p < 0.05, when compared with 75 μM. www.impactjournals.com/oncotarget 35211 Oncotarget degradative processes of tumor metastasis enzymatically and HOS cells, indicating the nobiletin-induced inhibition in U2OS and HOS cells. of MMP-2 and MMP-9 expressions at the transcriptional level (Figure 4A and 4B). Next, we assessed the single-site Nobiletin reverses EMT by increasing mutation plasmids of CREB and NF-κB and their promoter E-cadherin and decreasing Vimentin expressions activities. In U2OS cells, nobiletin at high concentrations in U2OS and HOS cells (75 and 100μM) seemed to inhibit MMP-2-CREB-mut but without significance (p = 0.062). Intriguingly, we observed To corroborate the anti-metastatic effect of nobiletin that a high concentration (100 μM) of nobiletin inhibited on human osteosarcoma cells, we examined the EMT- MMP-9-NF-κB-mut (p = 0.008) (Figure 4C and 4D). For related protein expression through western blotting. exploring whether nobiletin inhibites the binding activity High (100 μM) dose of nobiletin effectively inhibited the of NF-κB and CREB, the chromatin immunoprecipitation Vimentin (U2OS: p = 0.004; HOS: p < 0.001) expression, (ChIP) assay was performed. We found that 50 and 100 but dose-dependently increased the E-cadherin (U2OS: μM nobiletin reduced the activity of CREB (U2OS: p = p < 0.001; HOS: p < 0.001) expression (Supplementary 0.015, HOS: p < 0.001) and NF-κB (U2OS: p < 0.005, Figure S1). Accordingly, the nobiletin-induced expression HOS: p < 0.001) at the binding target sequence of MMP- is associated with EMT reversal. 2 and MMP-9 promoters, especially in HOS cells dose- dependently (Figure 4E). These results provided strong Nobiletin suppresses the activity of NF-κB, evidence that nobiletin prevents the binding activity of the CREB, and SP-1 of U2OS and HOS cells target sequence of CREB and NF-κB to down-regulate transcription levels of MMP-2 and MMP-9 in U2OS and To further define the role of the nuclear translocation HOS cells. of NF-κB and other transcription factors SP-1, CREB and AP-1, we examined the NF-κB expression in the Nobiletin suppresses cell migration and invasion cell cytosolic and nuclear fractions, and IKK dependent by down-regulating phosphorylation of ERK and pathway through western blotting. In U2OS and HOS JNK of U2OS and HOS cells cells, nobiletin considerably reduced the phosphorylation of p-IKKα/β (U2OS: p = 0.004; HOS: p = 0.001) and To assess whether nobiletin mediates p-IκBα (U2OS: p < 0.001; HOS: p = 0.002) (Figure phosphorylation of the three major mammalian MAPK 3A), and protein expression of NF-κB in the cell nuclear pathways, the effects of nobiletin on the constitutive (U2OS: p < 0.001; HOS: p = 0.003) fraction with the activation status of ERK, JNK, and p38 were analyzed concomitant increase of the NF-κB expression in the through western blotting. Nobiletin markedly reduced cytosolic (U2OS: p < 0.001; HOS: p = 0.007) fraction p-JNK (U2OS: p < 0.001; HOS: p < 0.001) and p-ERK (Figure 3B). Nobiletin down-regulated the p-CREB (U2OS: p < 0.001; HOS: p < 0.001); however, it did not (U2OS: p < 0.001; HOS: p < 0.001) and the SP-1 (U2OS: affect the p-p38 (U2OS: p = 0.227; HOS: p = 0.617) p < 0.001; HOS: p < 0.001) expressions in the nuclear activity in U2OS and HOS cells, and the protein levels fraction in U2OS and HOS cells, whereas nobiletin did not of total JNK, ERK, and p38 remained unchanged (Figure affect c-Jun (U2OS: p = 0.685; HOS: p = 0.909) and c-Fos 5A). To further investigate whether nobiletin mediates (U2OS: p = 0.885; HOS: p = 0.747) expressions (Figure the inhibition of cell invasion and migration as well as 3C and 3D). Furthermore, the inhibitory effect on nuclear MMP-2 and MMP-9 secretions mainly by inhibiting translocation of NF-κB from the cytoplasm was observed JNK and ERK phosphorylation, specific inhibitors of in U2OS and HOS cells using immunocytofluorescence ERK (U0126) and JNK (SP600125) were used in U2OS (Figure 3E). Thus, the transcriptional inhibition of MMP- cells. Zymography revealed that co-treatment with 2 and MMP-9 may down-regulate NF-κB and CREB, but inhibitors and nobiletin further reduced the proteolytic not AP-1. activities of MMP-2 and MMP-9 (Figure 5B). A similar effect was observed through western blotting (MMP- Nobiletin reduces promoter activities of MMP-2 2: p < 0.001; MMP-9: p < 0.001; p-ERK: p < 0.001; and MMP-9 of U2OS and HOS cells p-JNK: p < 0.001) in U2OS cells (Figure 5C and 5D). As shown in Figures 5E and Supplementary Figure S2, co- We further identify whether nobiletin suppresses the treatment with inhibitors and nobiletin further reduced transcription factors binding to the promoter of MMP-2 wound closure (p < 0.001), invasion (p < 0.001) and and MMP-9 to regulate their gene expressions in U2OS migration (p < 0.001) of U2OS cells compared with and HOS cells. The luciferase reporter assay revealed individual treatments, indicating that the inhibition of that nobiletin dose-dependently suppressed the promoter MMP-2 and MMP-9 to suppress U2OS cells motility, activities of MMP-2 (U2OS: p = 0.007; HOS: p = 0.002) invasion and migration was via down-regulation of the and MMP-9 (U2OS: p = 0.006; HOS: p = 0.006) in U2OS ERK and JNK pathways.

www.impactjournals.com/oncotarget 35212 Oncotarget Figure 3: NF-κB and CREB are crucial to nobiletin-mediated transcriptional inhibition of MMP-2 and MMP-9. A-D. U2OS and HOS cells were treated with nobiletin (0–100 μM) for 24 h. The nuclear fraction and total cell lysates were prepared as described in Materials and Methods. Levels of NF-κB, SP-1, c-Jun and c-Fos from nuclear (B and D) and levels of p-IKKα/β, IKKβ, p-IκBα, IκBα, CREB and p-CREB from the total cell lysates (A and C) were determined by western blotting. Quantitative results of p-IKKα/β, p-IκBα, NF-κB, SP-1, c-Jun, c-Fos and p-CREB protein levels were adjusted with β-actin, C23 or α-tubulin protein level. Concentration effects: p-IKKα/β (U2OS: F = 15.419, p = 0.004; HOS: F = 28.397, p = 0.001); p-IκBα (U2OS: F = 194.788, p < 0.001; HOS: F = 21.407, p = 0.002); total-p-CREB (U2OS: F = 61.097, p < 0.001; HOS: F = 34.873, p < 0.001); cytosol-NF-κB (U2OS: F = 60.412, p < 0.001; HOS: F = 16.628, p = 0.007); nuclear-NF-κB (U2OS: F = 50.038, p < 0.001; HOS: F = 17.562, p = 0.003); SP-1 (U2OS: F = 58.954, p < 0.001; HOS: F = 49.681, p < 0.001); c-Jun (U2OS: F = 0.402, p =0.685; HOS: F = 0.097, p =0.909); c-Fos (U2OS: F = 0.124, p =0.885; HOS: F = 0.207, p =0.747). aSignificantly different,p < 0.05, when compared with 0 μM. bSignificantly different,p < 0.05, when compared with 25 μM. E. U2OS and HOS cells were stained for NF-κB by immunofluorescence. Pre-treatment with 100 μM nobiletin for 24 h significantly inhibited the nuclear translocation of NF-κB (White arrows were added to direct against distinct changes). www.impactjournals.com/oncotarget 35213 Oncotarget Figure 4: Nobiletin inhibits MMP-2 and MMP-9 promoter and DNA-binding activity. A-D. U2OS and HOS cells were treated with nobiletin (0–100 μM) for 24 h and then subjected to luciferase assay to analyze the promoter activity of MMP-2 and MMP- 9. Upper: schematic of the wild type promoter region of the human MMP-2 and MMP-9 gene and the utilized mutant constructs. Lower: (A and B) U2OS and HOS cells were transfected with MMP-2 and MMP-9 promoter plasmid (wild type); (C and D) U2OS cells were transfected with MMP-2 (CREB-mut) and MMP-9 (NF-κB-mut) promoter plasmid. Transfected cells were treated with nobiletin (0–100 μM) for 24 h. Luciferaes activity was determined in triplicates, was normalized to the β-galactosidase. Concentration effects: promoter activity (wt-MMP-2 of U2OS: F = 13.634, p = 0.007 and HOS: F = 25.179, p = 0.002; wt-MMP-9 of U2OS: F = 14.477, p = 0.006 and HOS: F = 13.898, p = 0.006); CREB-mut MMP-2: F = 4.612, p = 0.062; NF-κB-mut MMP-9: F = 12.477, p = 0.008. aSignificantly different, p < 0.05, when compared with 0 μM. bSignificantly different, p < 0.05, when compared with 25 μM. E. ChIP assay was used to analyze the association of transcription factor CREB with MMP-2 and NF-κB with MMP-9 promoter region in U2OS and HOS cells. Concentration effects: CREB (MMP-2 of U2OS: F = 9.09, p = 0.015 and HOS: F = 72.564, p < 0.001); NF-κB (MMP-9 of U2OS: F = 14.909, p = 0.005 and HOS: F = 77.244, p < 0.001). aSignificantly different, p < 0.05, when compared with 0 μM. bSignificantly different, p < 0.05, when compared with 50 μM. www.impactjournals.com/oncotarget 35214 Oncotarget Figure 5: Effects of nobiletin on the MAPK pathway in U2OS and HOS cells. A. U2OS and HOS cells were treated with nobiletin (0–100 μM) for 24 h and the total cell lysates were then subjected to western blotting with anti-ERK, anti-JNK and anti-p38 antibodies. Quantitative results of ERK, JNK and p38 protein levels were adjusted to β-actin protein level. Concentration effects: p-ERK (U2OS: F = 22.386, p < 0.001 ;HOS: F = 25.074, p < 0.001); p-JNK (U2OS: F= 31.325, p < 0.001; HOS: F= 53.906, p < 0.001); p-p38 (U2OS: F = 1.694, p =0.227; HOS: F = 0.687, p =0.617). aSignificantly different, p < 0.05, when compared with control. bSignificantly different, p < 0.05, when compared with 25 μM. cSignificantly different, p < 0.05, when compared with 50 μM. B-E. U2OS cells were co- treated with specific protein inhibitor U0126 or SP600125, which incubated in the presence or absence of nobiletin (50 μM) for 24 h. (B) Gelatin zymography to analyze the activity of MMP-2 and MMP-9. (C and D) U2OS cells were co-treated with specific protein inhibitor U0126 or SP600125, which incubated in the presence or absence of nobiletin (50 μM) for 24 h and then the total cell lysates were subjected to western blotting with (C) anti-MMP-2, anti-MMP-9 and anti-ERK. MMP-2: F = 122.929, p < 0.001; MMP-9: F = 104.197, p < 0.001; p-ERK: F = 186.714, p < 0.001. (D) anti-MMP-2, anti-MMP-9 and anti-JNK. MMP-2: F = 56.897, p < 0.001; MMP-9: F = 50.368, p < 0.001; p-ERK: F = 26.765, p < 0.001. aSignificantly different,p < 0.05, when compared with 0 μM. bSignificantly different,p < 0.05, when compared with noibletin-treated group. cSignificantly different,p < 0.05, when compared with U0126 or SP600125 treated-group. (E) Cell migration and invasion ability of U2OS cells were measured by using a Boyden chamber assay. Migration: F = 68.12, p < 0.001; invasion: F = 102.991, p < 0.001. aSignificantly different, p < 0.05, when compared with 0 μM. bSignificantly different, p < 0.05, when compared with noibletin-treated group. cSignificantly different,p < 0.05, when compared with U0126 treated-group. dSignificantly different,p < 0.05, when compared with nobiletin plus U0126-treated group. eSignificantly different, p < 0.05, when compared with SP600125 treated-group. www.impactjournals.com/oncotarget 35215 Oncotarget DISCUSSION expressions of Bcl-2 and Cox-2 [18, 19, 23]. However, we found that nobiletin at low concentrations (up to 100 μM) The present study sought to determine the anti- without cytotoxicity dose-dependently reduced the cell metastatic effect of nobiletin on the human osteosarcoma motility, invasion and migration of U2OS and HOS cells. cells in vitro by monitoring the regulation of MMP- Similarly, such effect has been reported that nobiletin 2 and MMP-9 expressions and the possible signaling markedly reduces the migration ability of breast cancer pathways (Figure 6). The major findings were as follows: cells under noncytotoxic concentrations [26]. In general, (a) nobiletin at concentrations (up to 100 μM) without nobiletin is considered as a flavonoid with weak cytostatic cytotoxicity inhibited human osteosarcoma U2OS and properties (up to 100 μM) towards progression of cancer HOS cells motility, migration potential and invasiveness, cells in vivo and in vitro [22, 24, 27]. In the present study, accompanied by a decrease of MMP-2 and MMP-9 we used the concentration range 0-100 μM of nobiletin in expressions; (b) nobiletin reversed EMT by increasing all subsequent experiments. Moreover, pharmacokinetic E-cadherin and decreasing Vimentin expressions in U2OS study in rats for orally administered nobiletin (2 mg/ and HOS cells; (c) nobiletin suppressed the activities of kg) has been reported that after oral administration NF-κB, CREB, and SP-1 to reduce MMP-2 and MMP- of nobiletin, the Cmax of plasma concentration was 9 promoter activities in U2OS and HOS cells and (d) calculated to be 4.7 ± 0.5 ng/mL [28]. Therefore, more nobiletin suppressed cell migration and invasion by down- animal studies and clinical trials using the concentration regulating phosphorylation of ERK and JNK in U2OS and range of nobiletin are needed to further justify its clinical HOS cells. value. In human acute myeloid leukemia, gastric cancer During invasion, MMPs facilitate the degradation and hepatocellular carcinoma, nobiletin significantly and invasion of ECM components, and participate in suppresses cell proliferation and induces apoptosis the onset and progression of tumors [11, 29]. It is well- by inducing caspase signals and down-regulating the established that inhibitions of MMP gene expression or

Figure 6: Propose signaling pathways by which nobiletin inhibits cell motility, migration and invasion of human osteosarcoma cells. ⊝ indicates inhibition. www.impactjournals.com/oncotarget 35216 Oncotarget enzyme activity are early targets for preventing cancer κB. Moreover, nobiletin effectively suppressed SP-1 and metastasis [30]. Of all MMPs, MMP-2 and MMP-9 are CREB expressions, these findings are consistent with those most widely studied that both gain and loss-of function in reports on leukemia cells [51], cholangiocarcinoma are associated with osteosarcoma progression [31–34]. cells [52] and osteosarcoma cells [53], implicating the As reported by others about other flavonoids such as importance of NF-κB and CREB in the regulation of naringin and isorhamnetin, nobiletin could inhibit MMP- MMP-2 and MMP-9 in U2OS and HOS. Although c-Jun 2 and MMP-9 expressions as well as protease activities in and c-Fos have been thought to be indispensable for the U2OS and HOS cells to suppress migration and invasion inhibition of MMP gene expression and both require abilities [35, 36]. Because E-cadherin is regarded as a synergistic cooperation with either the NF-κB or the gatekeeper of the epithelial state in various epithelial cell SP-1 site, nobiletin had no effect on c-Jun and c-Fos in types and the inhibition of E-cadherin expression involves U2OS and HOS cells. It has been documented that NF- osteosarcoma invasion and metastasis, EMT is part of the κB is critically important for MMP-9 to mediate cancer process of cancer cell dissemination and transmigration cell invasion, and the binding of CREB to promoter is [37]. Interestingly, we observed that nobiletin reversed centrally involved in MMP-2 gene expression and related the EMT by suppressing Vimentin at a high concentration to the malignancy of cancer cells [54]. We, therefore, used (100 μM) and dose-dependently increasing E-cadherin mutated NF-κB and CREB reporter plasmid and further expressions. Previous study also reveals that nobiletin demonstrated that this inhibitory effect of nobiletin on effectively attenuated hypoxia-induced epithelial- MMP-2 and MMP-9 activity is partially reversed in mesenchymal transition of human H1299 lung cancer U2OS cells. Although the inhibitory effect on MMP-9 cells [38]. Besides, loss of cell–cell connections, epithelial promoter activity was found in high concentration (100 characteristics and remodeling of extracellular matrix μM) of nobiletin, we still supposed that the tendency to mediated by MMPs, has been acknowledged as a critical a decreased expression of SP-1 may explain the slight mechanism for cancer metastasis. Thus, EMT reversal decrease in the MMP-9 level. is considered as an efficient process of inhibiting tumor The MAPK pathways are prominent mechanisms metastasis [39]. Moreover, some researchers have shown that coordinate the transcription factors, such as NF-κB that EMT-related proteins have a close relationship and CREB, which are activated and phosphorylated by with MMP-2 and MMP-9 [40, 41], so we inferred that ERK and JNK kinase systems to regulate downstream nobiletin might inhibit the metastasis ability of U2OS MMP-2 and MMP-9 gene expression, subsequently and HOS cells by decreasing MMP-2 and MMP-9 protein increase tumor initiation and progression [55–57]. Review expressions and then the EMT-related proteins might be studies documented the importance of ERK and JNK also regulated by MMP-2 and MMP-9. Therefore, the protein, which phosphorylate multiple transcriptional down-regulation of MMP-2 and MMP-9 is related to the elements to promote cell proliferation, survival and inhibition of osteosarcoma cell metastasis and is partly invasion [58, 59]. Constitutive ERK activation increasing attributed to EMT reversal. cell survival and cell invasion is influenced by integrin, Since the activation of transcription factors to which binds and activates the MAPK to down-regulate the the promoter is primarily involved in the induction and E-cadherin adhesion molecule [60, 61]. However, our data maintenance of EMT and the diverse aspect of cancer demonstrated that nobiletin inhibited the phosphorylation biology, identification of these regulatory elements might of ERK and JNK, leading to the down-regulation of MMP- be manipulated to reduce MMPs expression for affecting 2 and MMP-9 expression and reversed EMT by increasing cancer metastasis [42–44]. Generally, a crucial c-Jun E-cadherin expression to decrease U2OS and HOS and c-Fos binding site, AP-1 is located in the promoter cellular migration and invasion. Previous studies shown region of inducible type of MMP genes. Blocking in the that nobiletin inhibits invasion and migration capacities AP-1 binding activity prevents tumor promoter-induced of nasopharyngeal carcinoma cell lines and glioma cell MMP-9 transcription [45]. Besides, NF-κB and SP-1 lines through down-regulating ERK and/ or JNK pathway transcription factors also play a co-activators role in the [62, 63]. Nobiletin also suppresses MMP-9 expression regulation of MMP-2 and MMP-9 [46, 47]. However, once by reducing p38 MAPK activity [64]. Furthermore, we the IKK complex phosphorylates serine residues of IκB, were interested in nobiletin’s inhibitory mechanism on resulting in NF-κB release and then translocates to the ERK and JNK signaling pathway that mediates MMP-2 nucleus [48]. Researchers have reported that IKK/NF-κB and MMP-9 expression, implicating in U2OS and HOS pathway implicated in cancer metastasis [49, 50]. Thus, cells metastasis. Co-treatment with nobiletin and U0126 we postulated that nobiletin mediated reduction of MMP-2 or SP600125 could further suppress MMP-2 and MMP-9 and MMP-9 expressions to inhibit U2OS and HOS cellular expressions and cell motility, invasion and migration in migration and invasion may be deeply related to the U2OS cells, suggesting that nobiletin inhibits U2OS and suppression of DNA-binding activity of these regulatory HOS cells metastasis through ERK and JNK pathways. factors. Consequently, nobiletin inhibited phosphorylation To summarize, this is the first study verifying of IKKα/β and IκBα and the nuclear translocation of NF- the effects of nobiletin on human osteosarcoma cell www.impactjournals.com/oncotarget 35217 Oncotarget metastasis. Nobiletin inhibits osteosarcoma cell metastasis Wound healing assay by suppressing the expression and binding activities of 5 NF-κB and CREB on MMP-2 and MMP-9 promoters. We plated 8×10 U2OS and HOS cells in 6 cm Furthermore, this inhibitory effect may down-regulate plates for 16 h and the cells were followed by starvation phosphorylation of ERK and JNK, reduce MMP-2 and overnight. Wounded by scratching with yellow tip, the MMP-9 expressions and partially reverse the EMT. cells were incubated with condition medium containing Thus, nobiletin is a promising candidate for preventing 0.5% FBS and then received various treatments of osteosarcoma progression; however, clarifying the cellular nobiletin (0, 25, 50, 75, 100 μM) for 0, 6, 12, 24 h. Cells and molecular mechanisms underlying tumor metastasis in were co-treated with specificity protein inhibitor U0126 the future may provide new therapeutic strategies. (ERK inhibitor) or SP600125 (JNK inhibitor) for 0, 24 h, and they were photographed using a phase-contrast MATERIALs AND METHODS microscope (Olympus CX41).

Cell culture and treatment Cell invasion and migration assay

Human osteogenic sarcoma U2OS (female; 15 years The tumor invasion assay was performed using old) and HOS (female; 13 years old) cells both obtained a modified Boyden chamber (Neuro Probe, Cabin from the Food Industry Research and Development John, MD) as described elsewhere [65]. Matrigel Institute (Hsinchu, Taiwan) were cultured in Dulbecco’s (Collaborative Biomedical Products, Bedford, MA) was Modified Eagle Medium and Eagle’s MEM (Gibco BRL, diluted with cold filtered distilled water to receive the Grand Island, NY, USA); mouse osteoblast MC3T3-E1 concentration 25 μg/50 μl and applied to the upper to cell was obtained from ATCC and cultured in Alpha 8 μm pore size polycarbonate membrane filters of the Modification of Eagle's Medium (Gibco BRL, Grand upper well. Briefly, U2OS and HOS cells were treated Island, NY, USA) supplemented with 10% fetal bovine with indicated concentrations (0, 25, 50, 75, 100 μM) serum (HyClone Laboratories, Inc. South Logan, Utah, of nobiletin for 24 h. The migrated cells passed through USA), penicillin (100 U/mL) and streptomycin (100 μg/ the basement membrane layer and clung to the bottom mL) in a humidified 37°C incubator with atmosphere of of the Boyden chamber membrane, while non-invasive 5% CO2. Adhering cells were detached by incubation with cells stayed in the upper chamber. The data were trypsin. After reaching confluence, the cells were treated presented as the average number of cells attached to with appropriate amounts of stock solution (0.1 M in the bottom surface from randomly chosen files under DMSO) of nobiletin (Enzo Life sciences. Ann Arbor, MI, a light microscope. In order to measure the ability of USA) and then incubated for indicated time periods. The osteosarcoma cells on migration, cells were seeded harvested culture medium and cells lysate were stored at into the upper well of Boyden chamber which were not -20°C until use. coated with Matrigel. Migration of cells treated with different concentrations of nobiletin was measured as described in the cell invasion assay. Analysis of cell viability [microculture tetrazolium (MTT) assay] Gelatin zymography To evaluate the cytotoxicity of nobiletin, a MTT MMP-2 and MMP-9 activities were evaluated [3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyl-tetrazolium by gelatin zymography. Osteosarcoma cells treated bromide] assay was performed to determine the cell with different concentrations (0, 25, 50, 75, 100 μM) viability. Briefly, U2OS and HOS cells were seeded at a of nobiletin for 24 h after plating 8×104 cells in 24-well density of 8×104 per well in a 24-well, and MC3T3-E1 cells plates for 16 h. The harvested culture medium (16 μl) were seeded at a density of 105 per well in a 96-well plate containing 10 μg of total protein was subjected to SDS- for 16 h. Then, cells were treated with nobiletin at indicated PAGE containing 0.1% gelatin. After electrophoresis, concentrations (0, 25, 50, 75, 100 μM) for additional 24 the gels was washed with washing buffer (2.5% Triton h. Each concentration was repeated three times. After the X100) to remove SDS and then incubated at 37°C in exposure period, the medium was removed and followed by reaction buffer (40 mM Tris-HCl, 10 mM CaCl , 0.01% washing the cells with PBS. Then, the medium was changed 2 NaN ). The proteolytic activity of MMP-2 and MMP- and incubated with MTT solution (0.5 mg/ml) (Sigma, St. 3 9 was performed by staining with Coomassie Brilliant Louis, MO, USA) for 4 h. After removing the medium, Blue R-250. The intensity of unstained bands on blue the viable cell number per dish was directly proportional background was measured by spot density measurement to the production of formazan which was solubilized using a densitometer (AlphaImager 2000, Alpha Innotech in isopropanol and measured spectrophotometrically at Corp, San Leandro, CA). 563 nm. The percentage of viable cells was estimated by comparing with the untreated control cells. www.impactjournals.com/oncotarget 35218 Oncotarget Western blot analysis out using ABI StepOne Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The relative gene After the treatment with different concentrations expressions of MMP-2 and MMP-9 were normalized to (0, 25, 50, 75, 100 μM) of nobiletin, western blotting glyceraldehydes 3-phosphate dehydrogenase (GAPDH). was conducted as described elsewhere [66]. The blot was subsequently operated with standard procedures and Immunocytofluorescence probed with the following primary antibodies: MMP- 2, MMP-9, ERK, JNK, phospho-ERK, phospho-JNK, Cells were cultured on glass coverslips and fixed phospho-p38, IKKβ, phospho-IKKα/β, IκBα, phospho- with 4% paraformaldehyde at room temperature for 20 min IκBα and phospho-CREB (Cell Signaling Technologies, followed by PBS washing for three times. The cells were Danvers, MA); vimentin, NF-κB, C23, SP-1, CREB, permeabilized with 0.5% Triton X-100 for 10 min. After c-Jun and c-Fos (Santa Cruz Biotechnology, California, washing with PBS, cells were blocked with 5% bovine USA); E-cadherin, p-38 and β-actin (BD biosciences, serum albumin in PBS for 1 h. Then the samples were Bedford, MA, USA). Protein expression was detected by incubated with rabbit-anti-NF-κB antibody (1:200) at 4°C chemiluminescence with an ECL detection kit (Millipore, overnight. Subsequently, the samples were then washed by Bedford, MA, USA). PBS and incubated with FITC-conjugated rabbit antibody for 1 h at room temperature. The nuclei were counterstained Preparation of nuclear fraction with DAPI for 5 min. The coverslips were then washed extensively and mounted on glass slides with mounting 5 We plated 6×10 U2OS and HOS cells in 6 cm medium (DAKO, Glostrup, Denmark) for quantitative plates for 16 h, followed by the treatment with different image analysis. The image of samples coverslips were concentrations (0, 25, 50, 75, 100 μM) of nobiletin for examined with Zeiss LSM 510 META confocal microscope. 24 h. Nuclear extracts were obtained as Sigma-Aldrich procedure described. The cells were lysed with ice-cold Luciferase reporter assay buffer A (10 mM Hepes pH 7.9, 1.5 mM MgCl2, 10 mM KCL, 1 mM DTT, 0.5 mM PMSF, plus protease and We plated 8×104 U2OS and HOS cells in 24-well phosphatase inhibitors), followed by dounce homogeniser plate for 16 h. Followed by plasmid DNA transfection. and microcentrifuge to shear the cytoplasmic membranes. pGL3-MMP-2 promoter (wild-type or CREB-mut) or Cytoplasmin extracts were on the supernatant, resuspended MMP-9 promoter (wild-type or NF-κB-mut) plasmid was pellet in cell lysis buffer (iNtRON Biotechnology, INC), transfected into U2OS and HOS cells using lipofectamine and pipetted up and down to achieve homogenous mix. 2000 (Invitrogen Life Technologies, Carlsbad, CA, Followed by centrifugation at 13000 rpm for 30 min at USA) according to the manufacturer’s instructions. After 4°C in a microcentrifuge, we harvested the nuclear extract incubation, cultured cells were treated with different from the supernatant and stored at -20°C. The total and concentrations (0, 25, 50, 75, 100 μM) of nobiletin for extracts protein were determined by BCA assay. 24 h. Cells were harvested and lysed by reporter lysis buffer. For quantification of luciferase and β-galactosidase RT-PCR and quantitative real-time PCR activity, cell lysates were subjected to luciferase assay using a Luciferase Assay system. Respectively, luciferase Total RNA was isolated from U2OS and HOS activity was normalized to β-galatosidase activity to cells using Total RNA Mini Kit (Geneaid Biotech account for the transfection efficiency. Ltd.), according to the manufacturer’s instructions. Total RNA (2 μg) was used to reverse transcribed Chromatin immunoprecipitation (ChIP) assay into 20 μl cDNA by ABI-High-Capacity RNA-to- cDNATM Kit (Applied Biosystems inc. Foster City, The CHIP assay procedure was performed California). The PCR was performed in mixture as described elsewhere [67]. Briefly, Protein-DNA containing 2 μl RT buffer, 2 μl RT random primer, 0.8 complexes were cross-linked using formaldehyde at a final μl dNTP mix (100 mM), 1 μl multi-scribeTM reverse concentration of 1% at room temperature for 10 min. Cells transcriptase. The appropriate primers (MMP-2, MMP- were washed twice with ice-cold PBS and were collected 9 and GAPDH) were used for amplification of these by centrifugation at 4°C, and resuspended in the cell genes, GAPDH was served as an internal control. The lysis buffer containing 50 mM Tris–HCl, pH 8, 10 mM following forward (F) primers, and reverse (R) primers EDTA, 1% SDS, and protease inhibitors. Cell lysates were were used: MMP-2-F: ggCATCCAggTTATCggggA, sonicated to give a DNA size of approximately 300 bp, R: ggCCCTgTCACTCCTgAgAT; MMP- and supernatants were diluted with dilution buffer (16.7 9-F: CgggTgTAgAgTCTCTCgCT, R: mM Tris–HCl, pH 8, 1% Triton X-100, 1.2 mM EDTA, CAACATCACCTATTggATCC; GAPDH-F: 167 mM NaCl, 0.01% SDS, and protease inhibitors). AgCCTTCTCCATggTTggTgAAgAC, R: Immune complexes were precleared with salmon sperm CggAgTCAACggATTTggTCgTAT. Q-PCR was carried DNA/protein G agarose slurry and then treated with www.impactjournals.com/oncotarget 35219 Oncotarget antibody overnight at 4 °C. DNA immunoprecipitated with 7. Wick W, Platten M and Weller M. Glioma cell invasion: antibodies specific to NF-κB, CREB or the control, rabbit regulation of metalloproteinase activity by TGF-beta. immunoblobulin G (IgG), was purified and extracted using Journal of neuro-oncology. 2001; 53:177-185. QIAquick PCR purification kit (Qiagen), according to the 8. Abecassis I, Olofsson B, Schmid M, Zalcman G and manufacturer’s instructions. The specific primer of NF- Karniguian A. RhoA induces MMP-9 expression at CD44 κB and CREB of human MMP-2 and MMP-9 promoter lamellipodial focal complexes and promotes HMEC-1 cell binding region were used for PCR amplification. invasion. Experimental cell research. 2003; 291:363-376. 9. O'Reilly MS, Wiederschain D, Stetler-Stevenson WG, Statistical analysis Folkman J and Moses MA. Regulation of angiostatin production by matrix metalloproteinase-2 in a model of The values are presented as the mean ± SE from concomitant resistance. The Journal of biological chemistry. three independent experiments. For all data statistical 1999; 274:29568-29571. analysis, the significant differences between the control group and the experiment group were performed using 10. Nissinen L and Kahari VM. Matrix metalloproteinases one-way ANOVA with post hoc Bonferroni test. Results in inflammation. Biochimica et biophysica acta. 2014; were considered to be statistically significant only if the 1840:2571-2580. p value < 0.05 (Sigma-Stat 2.0, Jandel Scientific, San 11. Li H, Zhang K, Liu LH, Ouyang Y, Bu J, Guo HB and Rafael, CA, USA). Xiao T. A systematic review of matrix metalloproteinase 9 as a biomarker of survival in patients with osteosarcoma. ACKNOWLEDGMENTS Tumour biology. 2014; 35:5487-5491. 12. Wu YJ, Neoh CA, Tsao CY, Su JH and Li HH. This work was supported by grants from National Sinulariolide Suppresses Human Hepatocellular Carcinoma Science Council, Executive Yuan, Taiwan (NSC 102- Cell Migration and Invasion by Inhibiting Matrix 2314-B-040 -006 -MY2). This study was also supported Metalloproteinase-2/-9 through MAPKs and PI3K/Akt by a grant from Chung Shan Medical University Hospital, Signaling Pathways. International journal of molecular Taiwan (CSH-2015-D-004). sciences. 2015; 16:16469-16482. 13. Lee AY, Fan CC, Chen YA, Cheng CW, Sung YJ, Hsu CONFLICTS OF INTEREST CP and Kao TY. Curcumin Inhibits Invasiveness and Epithelial-Mesenchymal Transition in Oral Squamous Cell The authors declare that no conflicts of interest exist. Carcinoma Through Reducing Matrix Metalloproteinase 2, 9 and Modulating p53-E-Cadherin Pathway. Integrative REFERENCES cancer therapies. 2015; 14:484-490. 14. Li L, Wang Y, Qi B, Yuan D, Dong S, Guo D, Zhang C and 1. Kager L, Potschger U and Bielack S. Review of mifamurtide Yu M. Suppression of PMA-induced tumor cell invasion in the treatment of patients with osteosarcoma. Therapeutics and migration by ginsenoside Rg1 via the inhibition of and clinical risk management. 2010; 6:279-286. NF-kappaB-dependent MMP-9 expression. Oncology reports. 2014; 32:1779-1786. 2. Mirabello L, Troisi RJ and Savage SA. Osteosarcoma incidence and survival rates from 1973 to 2004: data from 15. Zhang D, Ma QY, Hu HT and Zhang M. beta2-adrenergic the Surveillance, Epidemiology, and End Results Program. antagonists suppress pancreatic cancer cell invasion by Cancer. 2009; 115:1531-1543. inhibiting CREB, NFkappaB and AP-1. Cancer biology & therapy. 2010; 10:19-29. 3. Yadav L, Puri N, Rastogi V, Satpute P, Ahmad R and Kaur G. Matrix metalloproteinases and cancer - roles in threat 16. Ravindranath MH, Ramasamy V, Moon S, Ruiz C and and therapy. Asian Pacific journal of cancer prevention: Muthugounder S. Differential growth suppression of human APJCP. 2014; 15:1085-1091. melanoma cells by tea (Camellia sinensis) epicatechins (ECG, EGC and EGCG). Evidence-based complementary 4. Hsieh YS, Chu SC, Yang SF, Chen PN, Liu YC and Lu and alternative medicine: eCAM. 2009; 6:523-530. KH. Silibinin suppresses human osteosarcoma MG-63 cell invasion by inhibiting the ERK-dependent c-Jun/AP-1 17. Walle T. Methoxylated flavones, a superior cancer induction of MMP-2. Carcinogenesis. 2007; 28:977-987. chemopreventive flavonoid subclass? Seminars in cancer biology. 2007; 17:354-362. 5. Foroni C, Broggini M, Generali D and Damia G. Epithelial- mesenchymal transition and breast cancer: role, molecular 18. Moon JY, Cho M, Ahn KS and Cho SK. Nobiletin induces mechanisms and clinical impact. Cancer treatment reviews. apoptosis and potentiates the effects of the anticancer drug 2012; 38:689-697. 5-fluorouracil in p53-mutated SNU-16 human gastric cancer cells. Nutrition and cancer. 2013; 65:286-295. 6. Chang C and Werb Z. The many faces of metalloproteases: cell growth, invasion, angiogenesis and metastasis. Trends 19. Ma X, Jin S, Zhang Y, Wan L, Zhao Y and Zhou L. in cell biology. 2001; 11:S37-43. Inhibitory effects of nobiletin on hepatocellular carcinoma www.impactjournals.com/oncotarget 35220 Oncotarget in vitro and in vivo. Phytotherapy research: PTR. 2014; 31. Shang HS, Chang JB, Lin JH, Lin JP, Hsu SC, Liu CM, Liu 28:560-567. JY, Wu PP, Lu HF, Au MK and Chung JG. Deguelin inhibits 20. Chen J, Chen AY, Huang H, Ye X, Rollyson WD, Perry HE, the migration and invasion of U-2 OS human osteosarcoma Brown KC, Rojanasakul Y, Rankin GO, Dasgupta P and cells via the inhibition of matrix metalloproteinase-2/-9 in Chen YC. The flavonoid nobiletin inhibits tumor growth vitro. Molecules. 2014; 19:16588-16608. and angiogenesis of ovarian cancers via the Akt pathway. 32. Liao CL, Lin JH, Lien JC, Hsu SC, Chueh FS, Yu CC, Wu International journal of oncology. 2015; 46:2629-2638. PP, Huang YP, Lin JG and Chung JG. The crude extract of 21. Miyata Y, Sato T, Imada K, Dobashi A, Yano M and Ito Corni Fructus inhibits the migration and invasion of U-2 OS A. A citrus polymethoxyflavonoid, nobiletin, is a novel human osteosarcoma cells through the inhibition of matrix MEK inhibitor that exhibits antitumor metastasis in human metalloproteinase-2/-9 by MAPK signaling. Environmental fibrosarcoma HT-1080 cells. Biochemical and biophysical toxicology. 2015; 30:53-63. research communications. 2008; 366:168-173. 33. Kawashima A, Nakanishi I, Tsuchiya H, Roessner A, Obata 22. Lam KH, Alex D, Lam IK, Tsui SK, Yang ZF and Lee SM. K and Okada Y. Expression of matrix metalloproteinase 9 Nobiletin, a polymethoxylated flavonoid from citrus, shows (92-kDa gelatinase/type IV collagenase) induced by tumour anti-angiogenic activity in a zebrafish in vivo model and necrosis factor alpha correlates with metastatic ability in HUVEC in vitro model. Journal of cellular biochemistry. a human osteosarcoma cell line. Virchows Archiv. 1994; 2011; 112:3313-3321. 424:547-552. 23. Hsiao PC, Lee WJ, Yang SF, Tan P, Chen HY, Lee LM, 34. Liu SJ, Hu GF, Liu YJ, Liu SG, Gao H, Zhang CS, Wei Chang JL, Lai GM, Chow JM and Chien MH. Nobiletin YY, Xue Y and Lao WD. Effect of human osteopontin on suppresses the proliferation and induces apoptosis involving proliferation, transmigration and expression of MMP-2 and MAPKs and caspase-8/-9/-3 signals in human acute myeloid MMP-9 in osteosarcoma cells. Chinese medical journal. leukemia cells. Tumour biology. 2014; 35:11903-11911. 2004; 117:235-240. 24. Kawabata K, Murakami A and Ohigashi H. Nobiletin, a 35. Aroui S, Najlaoui F, Chtourou Y, Meunier AC, Laajimi A, citrus flavonoid, down-regulates matrix metalloproteinase-7 Kenani A and Fetoui H. Naringin inhibits the invasion and (matrilysin) expression in HT-29 human colorectal cancer migration of human glioblastoma cell via downregulation cells. Bioscience, biotechnology, and biochemistry. 2005; of MMP-2 and MMP-9 expression and inactivation of p38 69:307-314. signaling pathway. Tumour biology. 2015. 25. Shi MD, Liao YC, Shih YW and Tsai LY. Nobiletin 36. Li C, Yang D, Zhao Y, Qiu Y, Cao X, Yu Y, Guo H, attenuates metastasis via both ERK and PI3K/Akt pathways Gu X and Yin X. Inhibitory Effects of Isorhamnetin in HGF-treated liver cancer HepG2 cells. Phytomedicine. on the Invasion of Human Breast Carcinoma Cells by 2013; 20:743-752. Downregulating the Expression and Activity of Matrix 26. Baek SH, Kim SM, Nam D, Lee JH, Ahn KS, Choi SH, Metalloproteinase-2/9. Nutrition and cancer. 2015; Kim SH, Shim BS, Chang IM and Ahn KS. Antimetastatic 67:1191-1200. effect of nobiletin through the down-regulation of CXC 37. Yang H, Zhang Y, Zhou Z, Jiang X and Shen A. chemokine receptor type 4 and matrix metallopeptidase-9. Transcription factor Snai1-1 induces osteosarcoma invasion Pharmaceutical biology. 2012; 50:1210-1218. and metastasis by inhibiting E-cadherin expression. 27. Morley KL, Ferguson PJ and Koropatnick J. Tangeretin and Oncology letters. 2014; 8:193-197. nobiletin induce G1 cell cycle arrest but not apoptosis in 38. Gao XJ, Liu JW, Zhang QG, Zhang JJ, Xu HT and Liu human breast and colon cancer cells. Cancer letters. 2007; HJ. Nobiletin inhibited hypoxia-induced epithelial- 251:168-178. mesenchymal transition of lung cancer cells by inactivating 28. Onoue S, Nakamura T, Uchida A, Ogawa K, Yuminoki K, of Notch-1 signaling and switching on miR-200b. Die Hashimoto N, Hiza A, Tsukaguchi Y, Asakawa T, Kan T Pharmazie. 2015; 70:256-262. and Yamada S. Physicochemical and biopharmaceutical 39. Tsai JH and Yang J. Epithelial-mesenchymal plasticity characterization of amorphous solid dispersion of in carcinoma metastasis. Genes & development. 2013; nobiletin, a citrus polymethoxylated flavone, with 27:2192-2206. improved hepatoprotective effects. European journal of 40. Yoo YA, Kang MH, Lee HJ, Kim BH, Park JK, Kim HK, pharmaceutical sciences. 2013; 49:453-460. Kim JS and Oh SC. Sonic hedgehog pathway promotes 29. Willis AL, Sabeh F, Li XY and Weiss SJ. Extracellular metastasis and lymphangiogenesis via activation of Akt, matrix determinants and the regulation of cancer cell EMT, and MMP-9 pathway in gastric cancer. Cancer invasion stratagems. Journal of microscopy. 2013; research. 2011; 71:7061-7070. 251:250-260. 41. Robichaud N, del Rincon SV, Huor B, Alain T, Petruccelli 30. Zhang L, Shi J, Feng J, Klocker H, Lee C and Zhang J. LA, Hearnden J, Goncalves C, Grotegut S, Spruck CH, Type IV collagenase (matrix metalloproteinase-2 and -9) Furic L, Larsson O, Muller WJ, Miller WH and Sonenberg in prostate cancer. Prostate cancer and prostatic diseases. N. Phosphorylation of eIF4E promotes EMT and metastasis 2004; 7:327-332. www.impactjournals.com/oncotarget 35221 Oncotarget via translational control of SNAIL and MMP-3. Oncogene. 53. Yang SF, Lee WJ, Tan P, Tang CH, Hsiao M, Hsieh 2015; 34:2032-2042. FK and Chien MH. Upregulation of miR-328 and 42. Benbow U and Brinckerhoff CE. The AP-1 site and MMP inhibition of CREB-DNA-binding activity are critical gene regulation: what is all the fuss about? Matrix biology. for resveratrol-mediated suppression of matrix 1997; 15:519-526. metalloproteinase-2 and subsequent metastatic ability in 43. Shapiro SD. Matrix metalloproteinase degradation of human osteosarcomas. Oncotarget. 2015; 6:2736-2753. extracellular matrix: biological consequences. Current doi: 10.18632/oncotarget.3088. opinion in cell biology. 1998; 10:602-608. 54. Sun B, Rong R, Jiang H, Zhang H, Wang Y, Bai X, Zhang 44. Chien MH, Ying TH, Hsieh YS, Chang YC, Yeh CM, Ko M, Ma J, Xia S, Shu W, Zhang L and Leng J. Prostaglandin JL, Lee WS, Chang JH and Yang SF. Dioscorea nipponica E2 receptor EP1 phosphorylate CREB and mediates MMP2 Makino inhibits migration and invasion of human oral expression in human cholangiocarcinoma cells. Molecular cancer HSC-3 cells by transcriptional inhibition of matrix and cellular biochemistry. 2013; 378:195-203. metalloproteinase-2 through modulation of CREB and AP-1 55. Reddy KB, Nabha SM and Atanaskova N. Role of MAP activity. Food and chemical toxicology. 2012; 50:558-566. kinase in tumor progression and invasion. Cancer metastasis 45. Xia Y, Lian S, Khoi PN, Yoon HJ, Joo YE, Chay KO, reviews. 2003; 22:395-403. Kim KK and Do Jung Y. Chrysin inhibits tumor promoter- 56. Kim D, Kim S, Koh H, Yoon SO, Chung AS, Cho KS induced MMP-9 expression by blocking AP-1 via and Chung J. Akt/PKB promotes cancer cell invasion suppression of ERK and JNK pathways in gastric cancer via increased motility and metalloproteinase production. cells. PloS one. 2015; 10:e0124007. FASEB journal: official publication of the Federation 46. Wang IK, Lin-Shiau SY and Lin JK. Suppression of of American Societies for Experimental Biology. 2001; invasion and MMP-9 expression in NIH 3T3 and v-H- 15:1953-1962. Ras 3T3 fibroblasts by lovastatin through inhibition of ras 57. Chakraborti S, Mandal M, Das S, Mandal A and isoprenylation. Oncology. 2000; 59:245-254. Chakraborti T. Regulation of matrix metalloproteinases: 47. Qin H, Sun Y and Benveniste EN. The transcription an overview. Molecular and cellular biochemistry. 2003; factors Sp1, Sp3, and AP-2 are required for constitutive 253:269-285. matrix metalloproteinase-2 gene expression in astroglioma 58. Kohno M and Pouyssegur J. Targeting the ERK signaling cells. The Journal of biological chemistry. 1999; pathway in cancer therapy. Annals of medicine. 2006; 274:29130-29137. 38:200-211. 48. Karin M. NF-kappaB as a critical link between 59. Lei YY, Wang WJ, Mei JH and Wang CL. Mitogen- inflammation and cancer. Cold Spring Harbor perspectives activated protein kinase signal transduction in solid tumors. in biology. 2009; 1:a000141. Asian Pacific journal of cancer prevention: APJCP. 2014; 49. Song L, Liu D, Zhao Y, He J, Kang H, Dai Z, Wang X, 15:8539-8548. Zhang S and Zan Y. Sinomenine inhibits breast cancer 60. Du J, Miller AJ, Widlund HR, Horstmann MA, cell invasion and migration by suppressing NF-kappaB Ramaswamy S and Fisher DE. MLANA/MART1 and activation mediated by IL-4/miR-324-5p/CUEDC2 axis. SILV/PMEL17/GP100 are transcriptionally regulated by Biochemical and biophysical research communications. MITF in melanocytes and melanoma. The American journal 2015; 464:705-710. of pathology. 2003; 163:333-343. 50. Liu B, Sun L, Liu Q, Gong C, Yao Y, Lv X, Lin L, Yao H, 61. Woods D, Cherwinski H, Venetsanakos E, Bhat A, Gysin Su F, Li D, Zeng M and Song E. A cytoplasmic NF-kappaB S, Humbert M, Bray PF, Saylor VL and McMahon M. interacting long noncoding RNA blocks IkappaB Induction of beta3-integrin gene expression by sustained phosphorylation and suppresses breast cancer metastasis. activation of the Ras-regulated Raf-MEK-extracellular Cancer cell. 2015; 27:370-381. signal-regulated kinase signaling pathway. Molecular and 51. Chen YJ and Chang LS. Simvastatin induces NFkappaB/ cellular biology. 2001; 21:3192-3205. p65 down-regulation and JNK1/c-Jun/ATF-2 activation, 62. Chien SY, Hsieh MJ, Chen CJ, Yang SF and Chen MK. leading to matrix metalloproteinase-9 (MMP-9) but Nobiletin inhibits invasion and migration of human not MMP-2 down-regulation in human leukemia cells. nasopharyngeal carcinoma cell lines by involving ERK1/2 Biochemical pharmacology. 2014; 92:530-543. and transcriptional inhibition of MMP-2. Expert opinion on 52. Lee GR, Jang SH, Kim CJ, Kim AR, Yoon DJ, Park therapeutic targets. 2015; 19:307-320. NH and Han IS. Capsaicin suppresses the migration of 63. Lien LM, Wang MJ, Chen RJ, Chiu HC, Wu JL, Shen cholangiocarcinoma cells by down-regulating matrix MY, Chou DS, Sheu JR, Lin KH and Lu WJ. Nobiletin, metalloproteinase-9 expression via the AMPK-NF-kappaB a Polymethoxylated Flavone, Inhibits Glioma Cell Growth signaling pathway. Clinical & experimental metastasis. and Migration via Arresting Cell Cycle and Suppressing 2014; 31:897-907.

www.impactjournals.com/oncotarget 35222 Oncotarget MAPK and Akt Pathways. Phytotherapy research: PTR. for the British Industrial Biological Research Association. 2016; 30:214-221. 2010; 48:1052-1058. 64. Kim JJ, Korm S, Kim WS, Kim OS, Lee JS, Min HG, Chin 66. Yang SF, Yang WE, Chang HR, Chu SC and Hsieh YS. YW and Cha HJ. Nobiletin suppresses MMP-9 expression Luteolin induces apoptosis in oral squamous cancer cells. through modulation of p38 MAPK activity in human dermal Journal of dental research. 2008; 87:401-406. fibrobalsts. Biological & pharmaceutical bulletin. 2014; 67. Chen MW, Hua KT, Kao HJ, Chi CC, Wei LH, Johansson 37:158-163. G, Shiah SG, Chen PS, Jeng YM, Cheng TY, Lai TC, 65. Yang SF, Chen MK, Hsieh YS, Yang JS, Zavras AI, Hsieh Chang JS, Jan YH, Chien MH, Yang CJ, Huang MS, et al. YH, Su SC, Kao TY, Chen PN and Chu SC. Antimetastatic H3K9 histone methyltransferase G9a promotes lung cancer effects of Terminalia catappa L. on oral cancer via a invasion and metastasis by silencing the cell adhesion down-regulation of metastasis-associated proteases. Food molecule Ep-CAM. Cancer research. 2010; 70:7830-7840. and chemical toxicology: an international journal published